Developer Guide

Developer Guide for Intel® oneAPI Math Kernel Library Linux*

ID 766690
Date 3/22/2024
Public

A newer version of this document is available. Customers should click here to go to the newest version.

Document Table of Contents

Versions of the Intel® GPU Optimized HPCG

The Intel® GPU Optimized HPCG package includes prebuilt HPCG benchmark for Intel MPI 2021.11.0 or higher. All the files of the benchmark are located in the /share/mkl/benchmarks/hpcg/hpcg_gpu subdirectory of the Intel® oneAPI Math Kernel Library (oneMKL) directory. The following versions of the benchmark are available:

File in …/benchmarks/hpcg/hpcg_gpu/bin

Description

xhpcg_impi_pvc with libhpcg_kernels_impi_pvc.so

The Intel® GPU optimized version of the benchmark, optimized for Intel Data Center GPU Max Series family of discrete GPUs. The prebuilt binary and companion shared library are ahead of time (AOT) compiled for "-device pvc", so they may not work on other Intel GPUs.

hpcg.dat

HPCG benchmark input file.

The Intel Optimized HPCG package also includes the source code necessary to build these versions of the benchmark for other MPI implementations, such as MPICH. Build instructions are available in the README.md file included with the package.

Product and Performance Information

Performance varies by use, configuration and other factors. Learn more at www.Intel.com/PerformanceIndex.

Notice revision #20201201